At the Frontiers of Lung Fibrosis Therapy (Nature Biotechnology)

September, 2013.

Despite recent failures, a flurry of activity in biotech and pharma is raising hopes that an effective drug for this fatal lung disease is on the way. Ken Garber investigates.

Despite a new drug approved in Europe, Japan and Canada, pulmonologists have little to offer patients with idiopathic pulmonary fibrosis (IPF).  InterMune’s Esbriet (pirfenidone), approved by the European Medical Agency in March, 2011, continues to roll out in Europe and has been on the market in Japan since 2008. But Esbriet has shown limited and inconsistent efficacy in IPF, which in the U.S. kills as many people (40,000) as breast cancer. Most die within three years of diagnosis.  “Your lungs turn to styrofoam, basically,” says Suzanne Bruhn, CEO of Promedior, an IPF therapy company in Lexington, Massachusetts.

A clinical trial shocked the field in 2011 by showing that the standard of care at the time actually harmed patients.  As of now, “there is no specific drug therapy that is directed at fibrosis in the lung that we know is likely to be beneficial,” says Kevin Brown, a pulmonologist at National Jewish Health in Denver, Colorado. Five other treatments have failed late-stage trials since 2009. And yet, “the mood of the field is cautious optimism,” says Andrew Tager, a fibrosis researcher at Harvard Medical School in Boston, Massachusetts.  Several current drug targets are based on compelling science.  (Table 1.) “We’re doing trials on things that make sense, have very good scientific rationale, and have strong preclinical data,” says Tager.

But IPF drug development will remain risky until clinically validated drug targets are identified. In addition, showing a survival advantage requires long, expensive trials, while other endpoints may face regulatory hurdles.  The field badly needs an unambiguous success to stave off defeatism, offer realistic hope to patients, and ensure continued biotech and pharma investment.

Unstoppable scarring

 InterMune’s Esbriet is not that success.  Of the two completed phase 3 trials conducted in the U.S., Europe and Australia, only one showed a statistically significant difference in primary endpoint between the Esbriet and placebo arms 1.  The U.S. Food and Drug Administration (FDA) rejected Esbriet in 2010, mandating a new phase 3 trial, which is now nearing completion. “The majority of patients with IPF should not be treated with pirfenidone,” according to the most recent international evidence-based guidelines2. Esbriet, to the extent that it works in IPF, does so by an unknown mechanism of action.  That’s in sharp contrast to the new generation of experimental therapies, which build on a growing understanding of the pathogenesis of IPF.

Fibrosis, or excessive tissue scarring, is a wound healing response gone awry.  In IPF, some unknown factor injures the alveoli, which leads to an influx of fibroblasts and the conversion of local fibroblasts to myofibroblasts. Myofibroblasts, in turn, make and deposit collagen and other components of the extracellular matrix.  In normal wound healing, this provisional matrix gradually disappears as the tissue regains its normal structure and function.  But in fibrosis the scarring continues, amplified and perpetuated by several positive feedback loops.  There’s a growing consensus that IPF is a disease of disordered crosstalk between lung epithelial cells and mesenchymal cells like fibroblasts.  “There’s some dysfunction of the epithelium in response to injury that leads to the release of factors that activate the mesenchyme,” says Paul Noble, a fibrosis researcher at Cedars-Sinai Medical Center in Los Angeles.

Several of these factors have become drug targets.  {See Fig.]  But distinguishing the disease drivers from the bystanders is not easy.  The best validated driver is transforming growth factor β (TGFβ), a pleiotropic cytokine that is critical for maintaining a cell’s homeostasis.  In wound healing, TGFβ recruits fibroblasts and macrophages to the site of injury, directly induces the differentiation of fibroblasts into myofibroblasts, and promotes collagen synthesis.  Although TGFβ overproduction in human disease tissue was first reported over 20 years ago, it’s a problematic drug target because of its role in normal homeostasis.

More recently, researchers have shown that the biomechanical properties of the extracellular matrix itself take an active part in perpetuating IPF.  “As you develop fibrosis that matrix becomes stiff, but then that stiff matrix delivers signals that can maintain and amplify the fibrosis,” says Tager.  Matrix stiffness activates TGFβ that’s bound in inactive form to the matrix, and blocks the synthesis of prostaglandin-E, a potent inhibitor of fibrosis.

Targeting the matrix

 These findings boost prospects for a therapy originally developed at Arresto Biosciences in Palo Alto, California.  It targets the enzyme lysyl oxidase-like-2 (LOXL2), which catalyzes the crosslinking of collagen.  Arresto researchers hypothesized that interfering with crosslinking would make it easier to clear remodeled collagen from the wound.

“But what we actually saw in the models were much more dramatic effects,” says Victoria Smith, Arresto’s former vice president for R&D.  “The activated fibroblasts would be disappearing from the disease area, in the models.  And we also consistently observed downregulation of both TGFβ signaling and TGFβ production.”

The likely explanation: a blocked matrix-driven positive feedback loop.  Stiff matrix, by releasing active TGFβ and other factors, delivers survival signals to fibroblasts and promotes their differentiation to myofibroblasts, which make more collagen and further stiffen the matrix.   Harvard’s Tager believes that interrupting this loop by blocking LOXL2  may reverse the cycle, causing established fibrosis to regress.  LOXL2 inhibition lowers levels of multiple profibrotic factors, including LOXL2 itself.

Arresto developed a humanized monoclonal antibody, simtuzumab, that allosterically blocks LOXL2 activity.  It proved effective in animal models, while decreasing TGFβ signaling3.  In January 2011 Gilead Sciences in Foster City, California acquired Arresto for $225 million.  Simtuzumab has progressed to phase 2 in IPF.

Smith considers simtuzumab’s unusual allosteric mechanism an important feature of the therapy, because substrate doesn’t compete with the antibody for binding.  LOXL2 expression is mostly confined to the disease setting, so Gilead postulates an ample therapeutic window.   The only question is whether the animal results will translate to humans.  Smith stresses that the therapy should work regardless of the upstream factors driving collagen deposition and matrix remodeling, because such drivers may be less important in maintaining fibrosis.   “It’s important to tackle the fibrotic response directly,” says Smith, who is now senior director of biology at Gilead.  “Will it be sufficient?  You know, that’s a good question.”

The apoptosis paradox

Another fibrosis drug target with unexpected effects is the lysophosphatidic acid (LPA) receptor, LPA1.  It was long known that some substance in the injured lung attracts fibroblasts to the site of injury.  In 2008, Tager’s group identified LPA as the fibroblast chemoattractant4.  Amira Pharmaceuticals, a San Diego startup, took a small molecule LPA1 inhibitor into development.

Besides attracting fibroblasts to the wound site, LPA also affects cell proliferation and survival.  “If you have an injury to the epithelium, production of LPA can cause more of the epithelial cells to apoptose, so it can… exacerbate that injury,” says Tager.  At the same time, LPA inhibits fibroblast apoptosis, and drives fibroblast gene expression.  Among the induced genes is connective tissue growth factor, which promotes fibroblast proliferation.  LPA signaling through LPA1 receptors on endothelial cells also induces vascular leak, which contributes to fibrosis through blood coagulation and fibrin accumulation.  LPA “seems to have a hand in most of the processes that are involved in driving pulmonary fibrosis,” says Tager.

Bristol-Myers Squibb acquired Amira for $325 million in September, 2011, based on preclinical results that Dean Sheppard, a fibrosis researcher at the University of California, San Francisco, calls “reasonably exciting.”  Amira’s drug reversed fibrosis in a mouse model of established fibrosis and  reduced tissue damage, vascular leak and fibrotic cytokine production.  But LPA biology is still at an early stage.  No one knows what cell types produce LPA, whether is it just produced at the site of injury or systemically, and especially why LPA signaling paradoxically drives apoptosis in epithelial cells but inhibits it in fibroblasts.  “I don’t have a great answer,” says Tager.

Bristol-Myers has advanced the drug, BMS-986020, to phase 2.  Because, unlike earlier drugs taken into fibrosis trials, it was developed specifically for fibrosis, “it’s a really exciting next step” for the field, says Noble.  “The challenge to this approach… is it’s a very very specific target.”  In other words, targeting this single molecule may or may not be enough to make a difference in human IPF.

Growth factor blockade

 Targeting multiple effector molecules is the strategy behind nintedanib  (BIBF 1120), a multikinase inhibitor from Boehringer Ingelheim in Ingelheim, Germany.  Like almost everything else in fibrosis, nintedanib has a connection to TGFβ.  There is growing evidence that TGFβ activation in fibrosis leads to the release of growth factors that perpetuate the disease process.  PDGF and FGF, for example, potently promote fibroblast proliferation.   In 2007, Boehringer researchers showed that a nintedanib analogue could inhibit the development of lung fibrosis in a mouse model, and could block the differentiation of fibroblasts to myofibroblasts in vitro, raising the possibility that a kinase inhibitor could mimic the downstream effects of TGFβ inhibition.

Nintedanib mainly blocks the receptors for PDGF, FGF and VEGF.  A double-blinded, randomized, placebo-controlled phase 2 trial produced encouraging results 5.  The trial did not meet its primary endpoint, but it came close, and it did meet multiple secondary endpoints.  “The measures of efficacy in that study are compelling,” says Kevin Brown.  “It’s the totality of the evidence that makes it worth pursuing, rather than any single endpoint.”  Boehringer Ingelheim launched two phase 3 trials in April 2011, and results should be available in early 2014.

Side effects, mainly diarrhea, nausea and vomiting, were a problem in phase 2, with many patients dropping out. “Think of it as a mild form of chemotherapy,” says Brown.

The TGFβ balancing act

 STX-100, an antibody from Biogen Idec in Cambridge, Massachusetts , has a very specific target, the αvβ6 integrin.  But its real target is TGFβ.  In fibrosis, secreted TGFβ accumulates in the extracellular matrix in inactive form, as part of a latent complex.  The αvβ6 integrin, expressed on the surface of epithelial cells, releases active TGFβ from the latent complex, freeing it to bind receptors on fibroblasts and activate TGFβ receptor signaling that leads to the familiar pattern of myofibroblast differentiation, recruitment and collagen production, driving fibrosis.  Crucially, STX-100 is targeting “not TGFβ itself but the activation of TGFβ in the pathologic target tissue,”  explains Tager.

This antibody has been long in development.  Dean Sheppard’s group at UCSF cloned the αvβ6 subunit in the late 1980s.  Noting that αvβ6 knockout mice were protected from fibrosis, Sheppard initiated a collaboration with Biogen to develop an anti-αvβ6 antibody.  But when Biogen merged with Idec Pharmaceuticals in San Diego, California in 2003, the project was shelved.

A few Biogen Idec scientists working on the project departed, forming Stromedix in 2007 in Cambridge, Massachusetts to in-license the anti-αvβ6 antibody and move it forward.  Five years later, Biogen Idec changed its mind about STX-100 and bought Stromedix, for $75 million.  The drug is now in phase 2a.

STX-100 performs a delicate balancing act, inhibiting only some TGFβ.  The αvβ6 integrin is highly upregulated in IPF patient lungs, where it appears to be an important trigger for TGFβ activation.  But because αvβ6 is just one of many potential activators of latent TGFβ, the antibody doesn’t touch most normal TGFβ signaling.  “The beauty of this target is we’re not systemically [this is what she said—my mistake]inhibiting the TGFβ pathway,” says Shelia Violette, the STX-100 program executive at Biogen Idec.  “We’re only inhibiting αvβ6-mediated TGFβ activation.”

But will it make a difference in IPF?  In mouse models of fibrosis, says Violette, STX-100 does at least as well as general systemic inhibitors of TGFβ.  But mouse models have been poor predictors in IPF.  “I’m reasonably confident we can find a dose that’s safe,” says Sheppard.  “But whether it’s going to be effective, you’ve just got to do the study.”

Best endpoint: breath or death?

 Phase 3 presents other issues.  “The big problem we face is, unlike cancer, we’ve not had a drug that sort of ‘shrinks the tumor,’ if you will,” says Noble.  So far, the only efficacy signal has been slower disease progression.  Esbriet, for example, at best merely delayed the decline in lung function compared to placebo.  Demonstrating even that  takes time. What’s worse, the most convenient endpoint, change in forced vital capacity, or FVC, has not been validated as clinically meaningful, in the view of some prominent IPF researchers. FVC basically measures how much air you can force from your lungs.  “It is simply a measure of static pulmonary physiology,” says Kevin Brown.  Brown’s view, shared by others, is that all-cause mortality would be a much better endpoint, as it’s clear that death is clinically meaningful.

But very few IPF trials have set mortality or survival as the primary endpoint.  To do that in IPF, a relatively rare disease, “ it takes such a large number of patients, that you just actually can’t do it,” says Promedior’s Bruhn.  Noble agrees.  “It’s just not feasible… for drug companies to invest the resources for mortality trials,” he says.  Hence the choice of FVC as a primary endpoint.

But no evidence exists showing that stabilizing or improving FVC translates to improved survival or better quality of life.  (It’s only known that a declining FVC is associated with a shorter lifespan.) So that endpoint introduces some uncertainty into the clinical trial outcome from the outset.  There is a history, Brown notes, of surrogate trial endpoints giving false answers.  “Sometimes you take shortcuts,” he says, “and you get to where you want to go, but sometimes you end up lost.”

Noble argues that it’s intuitively obvious that stabilizing lung function in IPF will lead to fewer deaths.  In any case, keeping the lungs working can only benefit patients.  “It’s a good thing not to have your lung function fall,” he says.

The FDA, in response to a query from Nature Biotechnology, issued this statement. “Clinical trials in IPF should emphasize outcomes that are clinically meaningful to patients.  Examples… include mortality, time to lung transplantation, disease progression, hospitalizations, and exacerbations. Given that there are no FDA-approved therapies for this fatal disease, FDA is open to the development of new potential endpoints (e.g. patient reporting).” No mention of FVC, which is the primary endpoint for the ongoing Esbriet phase 3 trial.

Meanwhile, the IPF field continues to grope its way forward.  “There’s still a lot we don’t know,” says Noble.  (Box 1) Some researchers are moving towards personalized medicine using prognostic and predictive biomarkers, as in cancer.  A UK-based study, now underway, prospectively evaluates such biomarkers in IPF patients, and is discovering new associations.(6)  Tager favors this approach, because of heterogeneity both in the natural history of the disease and in the biological pathways driving it. To tailor therapies to individuals, Tager says, “we may need to figure [out] which pathways are driving which patients’ fibrosis.”  And pathway redundancy, he adds, implies combination therapy.

But complex treatments aren’t a given.  “If one can target a central pathway like TGFβ effectively, you might end up with a nice effect across a broad range of those patients,” says Violette.  Brown agrees this is a plausible scenario.  “Finer and finer IPF phenotypes start to exclude patient populations from treatment trials,” he worries.  He expects that the current clinical trials will yield something that he can finally give desperate IPF patients to slow the relentless scarring of their lungs.

Box 1: Does the immune system matter?

Until recently, inflammation was thought to be a major driver of IPF.  In 1991, a small trial showed a survival advantage for the combination of two anti-inflammatories: prednisone, a steroid, and azathioprine, an immunosuppressant.  The combination became standard treatment.  (Later N-acetylcysteine, an antioxidant precursor, was added.)  But a larger randomized trial of the triple therapy was halted early in 2011 when the treatment actually proved harmful, increasing the rates of hospitalization and death compared to placebo.(7)

This failure of immunosuppressive therapy, along with an absence of inflammatory cells at the leading edge of fibrosis in lung biopsies, argue against inflammation’s importance.  It may drive early IPF, but even that is controversial.

But macrophages, phagocytic cells involved in inflammation, may still be important in IPF, although there is more evidence for this in kidney and liver.  Dysregulation of so-called “M2a” macrophages, which appear to promote fibrosis, has been implicated in several fibrotic diseases.

Promedior is targeting macrophages with its drug candidate, PRM-151, the recombinant form of the endogenous human protein Pentraxin-2.  Pentraxin-2 simultaneously recognizes damage-associated molecular patterns and binds receptors on monocytes.  Monocytes can differentiate into macrophages or into fibrocytes, circulating mesenchymal cells that may be a source of myofibroblasts in fibrosis.  When Pentraxin-2 binds to a monocyte, “it actually changes its differentiation away from the M2 macrophages and the fibrocytes that lead you down that pathway of destructive fibrosis, into a phenotype which is a regulatory macrophage,” explains Promedior CEO Suzanne Bruhn.

Regulatory macrophages are anti-fibrotic, says Bruhn, because they secrete the anti-inflammatory cytokine IL-10 and matrix metalloproteinases, which degrade collagen and break down scars.  Biologists at Rice University in Houston, Texas, have shown that Pentraxin-2 prevents the differentiation of monocytes into fibrocytes, and reverses bleomycin-induced lung fibrosis in rodents.  Promedior licensed the technology in 2006, and PRM-151 began clinical testing in 2009.   Phase 1b results, presented at the American Thoracic Society international conference in Philadelphia in May 2013, were encouraging.(8)

Tager thinks this approach may work.  “If we could drive things towards a regulatory macrophage phenotype, that actually again may be able to promote resolution of established fibrosis,” he says.  But Noble is pessimistic.  “I think they’re studying an epiphenomenon,” he says.  “I don’t think it’s highly likely to be effective in IPF.”


1.    Noble, P.W. et al., Lancet 377: 1760-9 (2011)

2.    Raghu G. et al., Am. J. Resp. Crit. Care Med. 183: 788-824 (2011).

3. Barry-Hamilton V. et al., Nat. Med. 16:1009-1017 (2010)

4. Tager A.M. et al., Nat. Med. 14: 45-54 (2008)

5. Richeldi L. et al., New Engl. J. Med. 65: 1079-87 (2011)

6. Maher T.M., Eur. Respir. Rev. 22, 148-52 (2013)

7.  Idiopathic Pulmonary Fibrosis Clinical Research Network, New Engl. J. Med., 366: 1868-77 (2012).

8. Van Den Blink B. et al., American Thoracic Society meeting abstract A5707 (2013).

Table 1: Selected agents in development for idiopathic pulmonary fibrosis (IPF)

Company Agent Target/mechanism Stage
InterMune Esbriet (pirfenidone), small molecule Unknown; affects myofibroblast proliferation & collagen synthesis Approved  in Europe, Japan, Canada;  new phase 3 in progress
Boehringer Ingelheim Nintedanib (BIBF 1120), small molecule Targets VEGFR, PDGFR, FGFR kinase receptors Phase 3
Gilead Sciences (Arresto Biosciences) Simtuzumab (GS-6624), humanized mAb Targets LOXL2, which catalyzes collagen crosslinking. Phase 2
MedImmune(Gaithersberg, MD) Tralokinumab (human MAb) Targets IL-13 Phase 2
Bristol-Myers Squibb (Amira Pharmaceuticals) BMS-986020 (AM152), small molecule Targets lysophosphatidic acid receptor (LPA1) Phase 2
FibrogenS. San Francisco FG-3019 human MAb Targets connective tissue growth factor (CTGF) Phase 2
Biogen Idec (Stromedix) STX-100, humanized mAb Targets αvβ6 integrin Phase 2a

Figure caption:

In IPF, the normal wound healing response goes fatally wrong, as resident fibroblasts become collagen-secreting myofibroblasts that cause continuous scar formation in the lung.   Driving this process are a variety of molecules that have become drug targets in the disease.

Comments are closed.